大数据学习26：关于HBase zookeeper 启动过程的一次剖析：

最新推荐文章于 2024-08-31 08:49:17 发布

原创最新推荐文章于 2024-08-31 08:49:17 发布 · 1.1k 阅读

0 ·

CC 4.0 BY-SA版权

文章标签：

#hbase #zookeeper

大数据专栏收录该内容

40 篇文章

订阅专栏

本文针对HBase启动过程中ZooKeeper出现的SSH问题进行了深入剖析，通过源码分析定位到了缺失的配置项hbase.zookeeper.quorum，最终解决了启动时的异常行为。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

关于hbase zookeeper 启动过程的一次剖析：
在做完一些配置后

[root@hadoop002 conf]# vi hbase-site.xml 
<configuration>
  <property>
    <name>hbase.rootdir</name>
    <value>hdfs://hadoop002:8020/hbase</value>
  </property>
  <property>
    <name>hbase.zookeeper.property.dataDir</name>
    <value>/home/hbase/zookeeper</value>
  </property>
  <property>
    <name>hbase.cluster.distributed</name>
    <value>true</value>
  </property>
</configuration>

启动hbase，发现 zookeeper 启动的时候会先ssh 0.0.0.0 导致我要输入密码才能运行。
这明显是不合理的，如果在复杂环境可能会出错，而且这种情况体现的是配置不可控，即便能正常运行，也存在风险。

[root@hadoop002 bin]# start-hbase.sh 
localhost: starting zookeeper, logging to /opt/software/hbase-1.2.0-cdh5.7.0/bin/../logs/hbase-root-zookeeper-hadoop002.out
starting master, logging to /opt/software/hbase-1.2.0-cdh5.7.0/logs/hbase-root-master-hadoop002.out
Java HotSpot(TM) 64-Bit Server VM warning: ignoring option PermSize=128m; support was removed in 8.0
Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=128m; support was removed in 8.0
hadoop002: starting regionserver, logging to /opt/software/hbase-1.2.0-cdh5.7.0/bin/../logs/hbase-root-regionserver-hadoop002.out
hadoop002: Java HotSpot(TM) 64-Bit Server VM warning: ignoring option PermSize=128m; support was removed in 8.0
hadoop002: Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=128m; support was removed in 8.0

那么，是不是我漏了什么配置呢？
开始检查。
1、启动的时候使用了 start-hbase.sh 。

[root@hadoop002 bin]# cat start-hbase.sh

发现这样一段话：

if [ "$distMode" == 'false' ] 
then
  "$bin"/hbase-daemon.sh --config "${HBASE_CONF_DIR}" $commandToRun master $@
else
  "$bin"/hbase-daemons.sh --config "${HBASE_CONF_DIR}" $commandToRun zookeeper
  "$bin"/hbase-daemon.sh --config "${HBASE_CONF_DIR}" $commandToRun master 
  "$bin"/hbase-daemons.sh --config "${HBASE_CONF_DIR}" \
    --hosts "${HBASE_REGIONSERVERS}" $commandToRun regionserver
  "$bin"/hbase-daemons.sh --config "${HBASE_CONF_DIR}" \
    --hosts "${HBASE_BACKUP_MASTERS}" $commandToRun master-backup
fi

其中关于zookeeper的是 $bin"/hbase-daemons.sh --config "${HBASE_CONF_DIR}" $commandToRun zookeeper ，那么检查bin目录下是否有zookeeper文件？

检查 zookeepers

[root@hadoop002 bin]# cat zookeepers.sh

发现这样一段话：

if [ "$HBASE_MANAGES_ZK" = "true" ]; then
  hosts=`"$bin"/hbase org.apache.hadoop.hbase.zookeeper.ZKServerTool | grep '^ZK host:' | sed 's,^ZK host:,,'`
  cmd=$"${@// /\\ }"
  for zookeeper in $hosts; do
   ssh $HBASE_SSH_OPTS $zookeeper $cmd 2>&1 | sed "s/^/$zookeeper: /" &
   if [ "$HBASE_SLAVE_SLEEP" != "" ]; then
     sleep $HBASE_SLAVE_SLEEP
   fi
  done
fi

可见 ssh $HBASE_SSH_OPTS $zookeeper ，$HBASE_SSH_OPTS在shell中好像是空，$zookeeper 为 $hosts ，而$hosts 为 "$bin"/hbase org.apache.hadoop.hbase.zookeeper.ZKServerTool | grep '^ZK host:' | sed 's,^ZK host:,,'
手动执行

 /opt/software/hbase-1.2.0-cdh5.7.0/bin/hbase org.apache.hadoop.hbase.zookeeper.ZKServerTool
ZK host: localhost

可见，hbase 运行 ZKServerTool 类得到的结果进行字符串截断获得。

去源码中查看 ZKServerTool 类：

package org.apache.hadoop.hbase.zookeeper;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.HBaseConfiguration;
import org.apache.hadoop.hbase.HBaseInterfaceAudience;
import org.apache.hadoop.hbase.HConstants;
import org.apache.hadoop.hbase.ServerName;
import org.apache.yetus.audience.InterfaceAudience;
import java.util.LinkedList;
import java.util.List;
public class ZKServerTool {
  public static ServerName[] readZKNodes(Configuration conf) {
    List<ServerName> hosts = new LinkedList<>();
    String quorum = conf.get(HConstants.ZOOKEEPER_QUORUM, HConstants.LOCALHOST);

    String[] values = quorum.split(",");
    for (String value : values) {
      String[] parts = value.split(":");
      String host = parts[0];
      int port = HConstants.DEFAULT_ZOOKEPER_CLIENT_PORT;
      if (parts.length > 1) {
        port = Integer.parseInt(parts[1]);
      }
      hosts.add(ServerName.valueOf(host, port, -1));
    }
    return hosts.toArray(new ServerName[hosts.size()]);
  }

  /**
   * Run the tool.
   * @param args Command line arguments.
   */
  public static void main(String args[]) {
    for(ServerName server: readZKNodes(HBaseConfiguration.create())) {
      // bin/zookeeper.sh relies on the "ZK host" string for grepping which is case sensitive.
      System.out.println("ZK host: " + server.getHostname());
    }
  }
}

发现：System.out.println("ZK host: " + server.getHostname()); 为类的输出。
server.getHostname() 为值，继续检查，server 为 readZKNodes(HBaseConfiguration.create()) 的结果。
去readZKNodes类中查看入参，发现HConstants.ZOOKEEPER_QUORUM, HConstants.LOCALHOST这两个，并且 HConstants.ZOOKEEPER_QUORUM 在前面。

继续在源码中查看HConstants.ZOOKEEPER_QUORUM 和 HConstants.LOCALHOST 方法。
在hbase/hbase-common/src/main/java/org/apache/hadoop/hbase/HConstants.java中发现

/** Name of ZooKeeper quorum configuration parameter. */ public static final String ZOOKEEPER_QUORUM = "hbase.zookeeper.quorum"; /** Host name of the local machine */ public static final String LOCALHOST = "localhost";

这时候发现这个参数：hbase.zookeeper.quorum
去hbase官网 default configuration（http://hbase.apache.org/book.html#hbase_default_configurations）中检查这个参数 hbase.zookeeper.quorum

3.Configure ZooKeeper In reality, you should carefully consider your ZooKeeper configuration. You can find out more about configuring ZooKeeper in zookeeper section. This configuration will direct HBase to start and manage a ZooKeeper instance on each node of the cluster. On node-a, edit conf/hbase-site.xml and add the following properties. <property> <name>hbase.zookeeper.quorum</name> <value>node-a.example.com,node-b.example.com,node-c.example.com</value> </property> <property> <name>hbase.zookeeper.property.dataDir</name> <value>/usr/local/zookeeper</value> </property>

发现在之前配置的时候少配了一个参数 hbase.zookeeper.quorum 。
在 hbase-site.xml 添加后，配置为：

<configuration> <property> <name>hbase.rootdir</name> <value>hdfs://hadoop002:8020/hbase</value> </property> <property> <name>hbase.zookeeper.property.dataDir</name> <value>/home/hbase/zookeeper</value> </property> <property> <name>hbase.cluster.distributed</name> <value>true</value> </property> <property> <name>hbase.zookeeper.quorum</name> <value>hadoop002</value> </property> </configuration>

这时，重启hbase
[root@hadoop002 bin]# stop-hbase.sh
[root@hadoop002 bin]# start-hbase.sh
hadoop002: starting zookeeper, logging to /opt/software/hbase-1.2.0-cdh5.7.0/bin/../logs/hbase-root-zookeeper-hadoop002.out starting master, logging to /opt/software/hbase-1.2.0-cdh5.7.0/logs/hbase-root-master-hadoop002.out Java HotSpot(TM) 64-Bit Server VM warning: ignoring option PermSize=128m; support was removed in 8.0 Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=128m; support was removed in 8.0 hadoop002: starting regionserver, logging to /opt/software/hbase-1.2.0-cdh5.7.0/bin/../logs/hbase-root-regionserver-hadoop002.out hadoop002: Java HotSpot(TM) 64-Bit Server VM warning: ignoring option PermSize=128m; support was removed in 8.0 hadoop002: Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=128m; support was removed in 8.0

发现这时候，为我所指定的 hadoop002 而不是localhost了。

总结：
从这个问题来看，主要解决的问题不是解决某个配置，而是在于去剖析程序的运行机制，从源码中分析hadoop工作机制，方便后期排错检查。思路比较重要！